[OSM-dev] Sanity check please - ignore GPS-point duplicates intrack data

michael_j at email.de michael_j at email.de
Tue Jun 19 23:30:13 BST 2007


> User_1400 whoever that might be just took the top spot on the gpx leader board (from nowhere 2 days ago) with over 7 million points.

ok. I'll remove the gpx tracks I uploaded lately with together 7 million gpx points from the OSM database and regenerate the gpx-tracks again using the suggested remove duplicate GPS points feature. I guess they might reduce the number of GPX-points from 70 million to something between 700 thousand to 70 thousand GPX-points. This should enable the download these GPS-tracks for larger areas again.

This will solve the issue for me.

Still I would like to encourage, that the OSM GPX-upload function deletes duplicate GPS-points, which appear in direct sequence. This should be simple and prevents GPS-downloads getting too large because data of misconfigured GPS-receiver had been uploaded. Besides it prevent many superfluous GPS-points from being stored and downloaded from the database. It might even significantly reduce the database load during GPS-upload and thus increase the GPS-upload speed.


> -----Ursprüngliche Nachricht-----
> Von: "Andy Robinson" <Andy_J_Robinson at blueyonder.co.uk>
> Gesendet: 15.06.07 10:39:37
> An:  <dev at openstreetmap.org>
> Betreff: RE: [OSM-dev] Sanity check please - ignore GPS-point duplicates intrack data

> michael_j at email.de wrote:
> >Sent: 14 June 2007 11:31 PM
> >To: dev at openstreetmap.org
> >Subject: [OSM-dev] Sanity check please - ignore GPS-point duplicates
> >intrack data
> >
> >Hi,
> >
> >last week I have uploaded to OSM many gpx-files generated by my GPS-
> >receiver. After uploading I noticed, that my GPS-receiver had been in
> >Sports mode instead of Navigation mode and thus took GPS-points maybe 20
> >times  more often. This resulted in huge GPX-files.
> >
> >As a result now downloading the uploaded GPS-points with Josm of other than
> >tiny areas  becomes impossible.I have seen areas with only few GPS-tracks
> >taken, however containing about 1 GPS-point per 20 square meters of the
> >displayed area.
> >
> >Today I downloaded the bounding box
> >(8.284247155697123,51.675814750870174,8.344998008366737,51.727143022309406)
> >which is about 20 km x 20 km = 40 000 000 square meter. Until now Josm had
> >downloaded about 1.500 000 GPS-points corresponding (380-http pages) and is
> >still not finished.
> >
> >Here is the URL:
> >http://www.openstreetmap.org/api/0.4/trackpoints?bbox=8.284247155697123,51.
> >675814750870174,8.344998008366737,51.727143022309406&page=0
> >
> >When now looking at this track data it appears that very many GPS-points
> >are identical with their predecessor. As no time information is provided -
> >their information content has zero value.
> >
> snip
> There is also a method of ensuring that the original GPX file you upload to
> OSM is not over bloated with duplicate point data. I had this problem
> originally with my PDA which spits out two different NMEA sentences with the
> same location data in sequence. I fixed this problem by using gpsbabel with
> the track "merge" function enabled. I do this on the GPX file rather than on
> the original NMEA sentences file as for some reason that appears for me to
> be more reliable. Works a treat.
> Cheers,
> Andy
> Andy Robinson
> andy_j_robinson at blueyonder.co.uk

More information about the dev mailing list